Prognostic factors were examined by multivariate analysis after a recent follow-up of the 300 patients with Hodgkin's disease pathologic stage (PS) I or II treated with radiotherapy +/- adjuvant combination chemotherapy in the prospective randomized trial of the Danish National Hodgkin Study. Initial biopsy material was classified according to the Rye histopathologic classification, the grading and subclassification proposed by the British National Lymphoma Investigation (BNLI), and tumor cell concentration in sections. Tumor cell concentration as a prognostic factor turned out to be better than the other classifications. However, if macroscopic tumor burden was taken into account both tumor cell concentration and the other histopathologic classification systems lost their prognostic significance. Significantly, however, a combination of macroscopic tumor burden and tumor cell concentration, yielding an estimate of the total tumor cell burden, was even better than the macroscopic tumor burden as a prognostic factor. In conclusion, a simple tumor cell concentration count seems to be the most useful form of histopathologic subtyping for prognostic purposes in early stage Hodgkin's disease.